Post on 17-Jan-2023
A
tuapth©
K
1
t2fi
bpOatre
r(
1d
Available online at www.sciencedirect.com
Separation and Purification Technology 61 (2008) 375–383
Preparative separation of ketoprofen enantiomers: Choice of mobile phasecomposition and measurement of competitive adsorption isotherms
Antonio E. Ribeiro a, Nuno S. Graca a, Luıs S. Pais a,∗, Alırio E. Rodrigues b
a Laboratory of Separation and Reaction Engineering, School of Technology and Management, Braganca Polytechnic Institute,Campus de Santa Apolonia, Apartado 1134, 5301-857 Braganca, Portugal
b Laboratory of Separation and Reaction Engineering, Faculty of Engineering, University of Porto,Rua Dr. Roberto Frias s/n, 4200-465 Porto, Portugal
Received 16 May 2007; received in revised form 9 November 2007; accepted 21 November 2007
bstract
The present work intends to investigate how mobile phase composition influences the adsorption behavior of ketoprofen enantiomers (a nons-eroidal anti-inflammatory drug) on an amylose-based chiral stationary phase (Chiralpak AD). Three mobile phase compositions were studied: thesual 20% ethanol/80% n-hexane mixture and two pure mobile phases; methanol and ethanol. Pulse and breakthrough experiments under prepar-tive conditions were carried out in order to measure and test adsorption isotherms. The results obtained show that, for preparative separations,
ure ethanol is a better mobile phase than the usual 20% ethanol/80% n-hexane mixture: it allows higher solubility of the racemate, lower retentionimes, and also a higher selectivity at high enantiomer concentrations. These are aspects of crucial importance when the final goal is to achieveigh productivity preparative separations, as it is shown for the simulated moving bed (SMB) operation.2007 Elsevier B.V. All rights reserved.
parat
mb(df
fabrifit
eywords: Ketoprofen; Enantiomer separation; Mobile phase composition; Pre
. Introduction
Ketoprofen (R,S)2-(3-benzoylphenylpropionic acid) enan-iomers (Fig. 1) belong to a family of chemicals named-arylpropionic acids, or profens, an important sub-class of therequently prescribed and used drugs called nonsteroidal anti-nflammatory drugs (NSAIDs).
In the last years, preparative chiral chromatography hasecome a more and more important separation process for theurification of pharmaceuticals and other added-value products.ne reason chromatography is preferred is that the process
llows both high yields and purities of both enantiomers. Onhe other hand, this technique is applicable to a wide variety of
acemic mixtures, since chromatographic stationary phases fornantiomer separation are now available.Among the different analytical methods for chiral sepa-ation of profens, high-performance liquid chromatographyHPLC) using chiral stationary phases (CSPs) has been the
∗ Corresponding author. Tel.: +351 273 303 090; fax: +351 273 313 051.E-mail address: pais@ipb.pt (L.S. Pais).
attaappha
383-5866/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.seppur.2007.11.010
ive chromatography; Simulated moving bed
ost employed. Particularly, high chiral recognition is providedy using the phenylcarbamate derivatives of polysaccharidescellulose and amylose-based) as CSPs. The amylose 3,5-imethylphenylcarbamate (e.g. Chiralpak AD) is the most usedor the separation of profens racemates [1–3].
The optimization of chiral separations in these adsorbents isrequently a complex task that requires, at a preparative scale,careful selection of its operating conditions. In the case of
inary or multicomponent mixtures, an additional complexityesults from the competition between the different componentsn the interaction with the stationary phase. Therefore, one of therst steps of the preliminary study of a chromatographic separa-
ion process is the determination of the equilibrium competitivedsorption isotherms of the two enantiomers that will contributeo explain the retention mechanism and allow the prediction ofhe production rate recoveries and separation costs. Addition-lly, solubility of the racemate, selectivity and retention timesre separation parameters very sensitive to changes in mobile
hase composition. In fact, the use of continuous separationrocesses, such as simulated moving bed (SMB) technology,as achieved high throughputs when high feed concentrationsnd short cycle times were applied [4].376 A.E. Ribeiro et al. / Separation and Purifica
tacpoiope
2
2
awi1
stiw(amwtoS
tB(w
2
gtpouom3s
a(c
S
2t
tmwocvaT
C
wrcwtohet(htfme
ccaemt2caa
3
3
i
Fig. 1. Chemical structure of ketoprofen enantiomers.
The aim of this work is the measurement of adsorp-ion equilibrium data of ketoprofen enantiomers, using thedsorption–desorption method, under different mobile phaseompositions. Additionally, solubility measurements of keto-rofen enantiomers and pulse and breakthrough experimentsn the different mobile phases were also performed. Model-ng of adsorption data and simulation of fixed-bed and SMBperation were carried out to justify the choice of the mobilehase composition for the preparative separation of ketoprofennantiomers.
. Experimental
.1. Equipment and materials
All the analysis were performed on a Jasco HPLC system withn UV-1575 multiwavelength detector set at 260 nm, equippedith a preparative cell (1.0 mm). A manual Rheodyne 7725(i)
njection valve was used with three different loops: 20 �L,00 �L and 1 mL.
Two chiral chromatographic columns were used with theame adsorbent material (Chiralpak AD, Daicel Chemical Indus-ries Ltd., Japan) and the same dimensions (250 mm L × 4,6 mm.d.). These two columns have different particle size: one column,ith a particle size of 10 �m, was used for analytical purposes
measurement of enantiomers concentrations); the other, withparticle size of 20 �m, was used in the preparative chro-atographic experiments (adsorption–desorption steps, pulsesith high concentrations and high injection volumes, and break-
hrough experiments). It must be pointed out that a particle sizef 20 �m is normally used for preparative separations, includingMB operation.
All isotherm measurements were performed at a constantemperature of 20 ◦C. Methanol, ethanol and n-hexane (Fluka,uchs, Switzerland) were all HPLC grade. Trifluoroacetic acid
TFA) was spectrophotometric grade, and racemic ketoprofenas of analytical grade, purchased from Merck (Madrid, Spain).
.2. Solubility measurements
The solubility measurements were carried out by using aravimetric method [5]. It consists in preparing saturated solu-ions of racemic ketoprofen in different solvents, which arelaced in a thermostatic water bath at a constant temperaturef 25 ◦C. After equilibration, a volume of 5 cm3 of the clear sat-rated solution is transferred to a previously weighed glass vial
f mass mV. The mass of the vial plus the saturated solution,VS, is then measured. After, the vial is placed in an oven at0 ◦C for solvent evaporation, until the remaining mass of theample does not change with time. Then, the final mass of vialtpfm
tion Technology 61 (2008) 375–383
nd residue, mVR, is measured, and solubility, expressed in massg) of solute per unit mass (kg) of solvent (on a solute-free basis)an be determined by using the following equation:
= (mVR − mV)
(mVS − mVR)× 103 (1)
.3. Determination of competitive adsorption isotherms byhe adsorption–desorption method
The experimental determination of the competitive adsorp-ion isotherms was carried out using the adsorption–desorption
ethod. In this method, the preparative column is saturatedith a large amount of feed solution with known concentrationf the two enantiomers, CF
i , until equilibrium is achieved. Theolumn is then completely regenerated with eluent. The elutedolume, resulting from this desorption step, is collected andnalyzed, in order to measure each enantiomer concentration.he mass balance
di V
d = εVcCFi + (1 − ε)Vcq
∗i (2)
ill allow to evaluate the concentration of each componentetained in the particle, q∗
i , in equilibrium with the feed con-entration, CF
i . In fact, q∗i is an overall retained concentration,
hich includes both the adsorbed material and the material inhe fluid inside pores. This is consistent with the simulationf the chromatographic process, considering a model based onomogeneous particles. In Eq. (2), Cd
i is the concentration ofach component in the eluted solution collected in the desorp-ion step, Vd is the eluted volume, Vc is the column volumeVc = 4.15 mL), and ε is the external bed porosity (considering aomogeneous particle model; ε = 0.4). This procedure will allowhe determination of a unique point of the adsorption isothermor each component (CF
i , q∗i ). The entire adsorption isotherm
easurement will require a set of adsorption–desorptionxperiments, using different feed concentrations.
In this work, several adsorption isotherm measurements werearried out, using racemic ketoprofen solutions at different con-entrations and using three different mobile phase compositions:20% ethanol/80% n-hexane mixture and two pure eluents;
thanol and methanol. All eluents include 0.01% of an acidicodifier (trifluoroacetic acid, TFA). For example, 1000 mL of
he 20% ethanol/80% n-hexane mixture is prepared adding00 mL ethanol, 800 mL n-hexane and 100 �L of TFA. The con-entration of each ketoprofen enantiomer in the feed (racemic)nd eluted solutions was evaluated by HPLC, equipped with thenalytical column described before.
. Modeling
.1. Modeling of competitive adsorption isotherms
After the experimental evaluation of the adsorption data, ansotherm model must be proposed in order to allow the simula-
ion of the adsorption behavior and the overall chromatographicrocess. The binary Langmuir model is usually a common choiceor this purpose (subscripts 1 and 2 represent the less and theore retained enantiomers, respectively):A.E. Ribeiro et al. / Separation and Purification Technology 61 (2008) 375–383 377
Table 1Model equations for fixed-bed chromatography (breakthrough), using the lineardriving force model
Mass balance equations∂Ci∂θ
= 1Pe
∂2Ci
∂x2 − ∂Ci∂x
− 1−εε
St(q∗i − qi)
∂qi
∂θ= St(q∗
i − qi) with i = 1 and 2 (component)
Equilibrium isothermsq∗
i = fi(C1, C2)
Initial conditionsθ = 0, ∀x, Ci = qi = 0
Boundary conditionsx = 0, Ci − 1
PedCidx
= CFi with CF
i known feed concentrationx = 1,
dCidx
= 0
x = z/L, dimensionless axial coordinate; θ = t/τ, dimensionless time variable;Pe = uiLc/Dax, Peclet number; ε, external bed porosity; k, mass transfer coef-ficient (s−1); τ=Lc/ui, holdup time (s); St = kτ, massic Stanton number.
Table 2Solubility (S) of racemic ketoprofen in the three solvent compositions(T = 25 ◦C), expressed in mass (g) of solute per unit mass (kg) of solvent (on asolute-free basis)
Solvent composition Solubility (S)
20% ethanol/80% n-hexane 101.311
q
Hokicw
Fso
csa
TE
M
2
1
1
Mma
00% ethanol 836.900% methanol 1463
∗1 = Qb1C1
1 + b1C1 + b2C2; q∗
2 = Qb2C2
1 + b1C1 + b2C2(3)
owever, this Langmuir model usually fails in the predictionf the chromatographic enantioseparation process. It is well
nown that, for the generality of chiral systems, the selectiv-ty factor, α, decreases with the increase of the chiral speciesoncentrations, which is not assumed by the Langmuir model,here α is constant: α = (q∗2/C2)/(q∗1/C1) = b2/b1. To over-
able 3stimated model parameters for ketoprofen adsorption isotherms for the three mobile
odel M N m1 m2 QA QB b1
0% ethanol/80% n-hexaneLG3
22
3 – – 137.3 1.72LLG4 4 0.6575 52.19 3.58LLG5 5 0.6148 0.3772 69.50 2.50BLG6 6 – – 8.355 159.3 5.19
00% ethanolLG3
22
3 – – 162.7 6.20LLG4 4 0.4788 38.72 1.39LLG5 5 0.5469 0.5931 24.75 1.98BLG6 6 – – 0.5777 192.2 2.23
00% methanolLG3
22
3 – – 177.9 5.70LLG4 4 1.366 × 10−5 177.2 5.73LLG5 5 7.024 × 10−6 7.210 × 10−8 177.2 5.73BLG6 6 – – 87.20 90.78 1.31
is the number of experimental points, N the number of estimated parameters, m1, models: LG3, Langmuir; LLG4, linear + Langmuir (m1 = m2); LLG5, linear + Langm
nd S.D. is the standard deviation.
ig. 2. Effect of the alcoholic content (ethanol/n-hexane-based solvents) in theolubility (S) of racemic ketoprofen enantiomers at 25 ◦C, expressed in mass (g)f solute per unit mass (kg) of solvent (on a solute-free basis).
ome this limitation, more complex models are usually used,uch as, the linear + Langmuir competitive isotherms (4) and (5)nd the bi-Langmuir competitive model (6):
q∗1 = mC1 + Qb1C1
1 + b1C1 + b2C2;
q∗2 = mC2 + Qb2C2
1 + b1C1 + b2C2(4)
q∗1 = m1C1 + Qb1C1
1 + b1C1 + b2C2;
q∗2 = m2C2 + Qb2C2
1 + b1C1 + b2C2(5)
q∗1 = QAb1C1
1 + b1C1 + b2C2+ QBb3C1
1 + b3C1 + b4C2;
q∗2 = QAb2C2
1 + b1C1 + b2C2+ QBb4C2
1 + b3C1 + b4C2(6)
phase compositions
b2 b3 b4 SQ S.D.
2 × 10−2 1.875 × 10−2 – – 2.180 0.33886 × 10−2 4.152 × 10−2 – – 1.493 0.28808 × 10−2 3.518 × 10−2 – – 0.6668 0.19810 × 10−2 1.479 × 10−1 1.213 × 10−2 1.149 × 10−2 0.5920 0.1924
9 × 10−3 7.278 × 10−3 – – 0.2222 0.10812 × 10−2 1.951 × 10−2 – – 0.1650 0.095746 × 10−2 2.656 × 10−2 – – 0.1558 0.095730 × 10−1 6.595 × 10−1 5.015 × 10−3 5.771 × 10−3 0.1291 0.08982
7 × 10−3 6.187 × 10−3 – – 0.4640 0.15634 × 10−3 6.216 × 10−3 – – 0.4640 0.16064 × 10−3 6.216 × 10−3 – – 0.4641 0.16525 × 10−4 1.195 × 10−2 1.103 × 10−2 6.707 × 10−4 0.4632 0.1701
2, QA, QB, b1, b2, b3 and b4 are the estimated parameters of the four isothermuir (m1 �= m2); BLG6, bi-Langmuir; SQ is the sum of squares of the residues
3 urifica
Cs
3
bfifa
Fde02
c
4
4
78 A.E. Ribeiro et al. / Separation and P
ontrary to the Langmuir isotherm (3), for these three models,electivity is a concentration-dependent function.
.2. Breakthrough simulations
In non-linear preparative chromatography, the modeling of
and profiles can be done using the linear driving force model forxed-bed chromatography. Table 1 presents the model equationsor binary breakthrough experiments, including the mass bal-nce equations, equilibrium isotherms, and initial and boundaryig. 3. Experimental elution profiles of ketoprofen enantiomers in the threeifferent mobile phase compositions: (a) 20% ethanol/80% n-hexane; (b) 100%thanol; (c) 100% methanol. Racemic ketoprofen concentrations in the range of.05–8.0 g/L; preparative column (particle diameter of 20 �m); UV detection at60 nm; flow rate of 1 mL/min; injected volume of 100 �L.
iaoh
Fde02
tion Technology 61 (2008) 375–383
onditions.
. Results and discussion
.1. Solubility measurements
The solubility of racemic ketoprofen mixtures was measured
n the three different solvents: 100% methanol, 100% ethanolnd 20% ethanol/80% n-hexane. Additionally, the dependencef ketoprofen solubility on the alcoholic content of an ethanol/n-exane-based solvent was also studied.ig. 4. Experimental elution profiles of ketoprofen enantiomers in the threeifferent mobile phase compositions: (a) 20% ethanol/80% n-hexane; (b) 100%thanol; (c) 100% methanol. Racemic ketoprofen concentrations in the range of.05–4.0 g/L; preparative column (particle diameter of 20 �m); UV detection at60 nm; flow rate of 1 mL/min; injected volume of 1 mL.
urifica
sThma
Flbmccl
cr
A.E. Ribeiro et al. / Separation and P
Table 2 shows that ketoprofen enantiomers have increasingolubilities for 20% ethanol, pure ethanol and pure methanol.
hese results also confirm that racemic drugs have considerablyigher solubilities in alcoholic solvents than in the traditionalobile phases used in analytical chiral separation, consisting inn alcohol–hydrocarbon combination, with high hydrocarbon
ig. 5. Comparison between model and experimental results for the equi-ibrium adsorption isotherms: (a) 20% ethanol/80% n-hexane mobile phase,i-Langmuir model (BLG6); (b) 100% ethanol mobile phase, bi-Langmuirodel (BLG6); (c) 100% methanol mobile phase, Langmuir model (LG3). Open
ircles for experimental concentration of the less retained enantiomer; closedircles for experimental concentration of the more retained enantiomer; solidines for adsorption isotherm model.
sbwue8nte
ppmwr
4
tcepi1uasikFethgnt
Fi1
tion Technology 61 (2008) 375–383 379
ontent [6]. This conclusion is underlined with the experimentalesults obtained for the dependency of the racemic ketoprofenolubility on the alcoholic content of an ethanol/n-hexane-ased solvent, shown in Fig. 2. The ketoprofen enantiomers,hich are insoluble in pure n-hexane, present increasing sol-bilities with the increase of the ethanol content. For purethanol, the solubility of racemic ketoprofen enantiomers is36.9 g/kg solvent. This result is in the same order of mag-itude of the ones obtained by Gracin and Rasmunson forhe solubility of ibuprofen enantiomers on pure methanol andthanol [7].
These results clearly show the importance of using a mobilehase composition with a high alcoholic content, since, forreparative scale separations, the high solubility of the race-ate is a major concern. The use of pure solvents will also beelcome at a preparative scale, because of the simplicity of their
eutilization [8].
.2. Pulse experiments
In order to have a global overview of the ketoprofen selec-ivity in the three different solvents and under preparativeonditions (high concentrations), a set of preliminary pulsexperiments was done. These experiments consisted of severalulse injections of different racemic ketoprofen concentrationsn the preparative column (particle size of 20 �m) and using the00 �L and 1 mL injection loops. A flow rate of 1 mL/min wassed. Results obtained are shown in Fig. 3 (for 100 �L loop)nd Fig. 4 (for 1 mL loop). For all pulse experiments it can beeen that the increase of the amount injected leads to a decreasen the retention time of both enantiomers, which is a well-nown behavior for systems described by favorable isotherms.igs. 3 and 4 show that the hydrocarbon mobile phase (20%thanol/80% n-hexane) presents considerable higher retentionimes than the pure mobile phases (ethanol and methanol). The
ydrocarbon mobile phase also leads to important chromato-raphic tails (see Fig. 4a), which is an indication of strongon-linear behavior and not welcome for preparative separa-ions. Comparing the results obtained for the two pure alcoholig. 6. Comparison between model (lines) and experimental (points) selectiv-ties for racemic mixtures. Squares for 20% etanol/80% n-hexane, circles for00% ethanol, and triangles for 100% methanol mobile phase.
380 A.E. Ribeiro et al. / Separation and Purification Technology 61 (2008) 375–383
Fig. 7. Saturation (adsorption) and regeneration (desorption) curves for total feed concentrations of 10, 20, 30, and 40 g/L (racemic mixtures). Comparison betweenexperimental (points) and simulation (lines) results. Mobile phase: 100% ethanol; flow rate: 0.5 mL/min. Model parameters used: ε = 0.4, Pe = 3500, k = 5.0 s−1
(St = kτ = 1000), and bi-Langmuir (BLG6) model parameters (see Table 3).
urifica
mcdaf
4m
iap(L
st2atbobr
fdctpdateeumplhie
4
Fr(
A.E. Ribeiro et al. / Separation and P
obile phases (Fig. 3b and c for 100 �L and Fig. 4b andfor 1 mL injection loops) it can be clearly concluded that,
espite higher ketoprofen solubility, pure methanol does notllows acceptable selectivity values and, consequently, ketopro-en enantioseparation.
.3. Multicomponent adsorption isotherm experiments andodeling
The experimental determination of competitive adsorptionsotherms for ketoprofen enantiomers was carried out using thedsorption–desorption method for the three mobile phase com-ositions and fitted to the four isotherm models presented beforeEqs. (3)–(6)). The isotherm parameters were estimated using aevenberg–Marquardt algorithm.
Table 3 presents the numerical results obtained and Fig. 5hows the good agreement between model and experimen-al results for the equilibrium adsorption isotherms. For the0% ethanol/80% n-hexane and 100% ethanol mobile phases,lthough the linear + Langmuir model reasonably describeshe adsorption behavior, more complex models, such as the
i-Langmuir isotherm, better simulate the experimental databtained. For the 100% methanol mobile phase, the adsorptionehavior is well described by the Langmuir model and no betteresults are obtained with more complex models (see Table 3).soo
ig. 8. SMB separation regions under negligible mass transfer resistances for the thange); (b) 5; (c) 10; (d) 20 g/L. The separation regions for 20% ethanol/80% n-hexanbigger separation regions) and pure methanol (smaller separation regions) are locate
tion Technology 61 (2008) 375–383 381
Fig. 6 compares the experimental and model selectivitiesor the three mobile phase compositions and illustrates threeifferent scenarios. For 100% methanol, selectivity is low andonstant, which means that the separation of ketoprofen enan-iomers hardly can be achieved using pure methanol as mobilehase. The common 20% ethanol/80% n-hexane mobile phase,espite its high selectivity for low concentrations, presentsstrong decrease in selectivity with the increase of enan-
iomers concentrations. The better situation is obtained for 100%thanol, where selectivity maintains high values even for highnantiomer concentrations. In conclusion, pure ethanol can besed for ketoprofen enantioseparation, presenting better perfor-ances than the common 20% ethanol/80% n-hexane mobile
hase: it allows significantly higher enantiomer solubilities,ower retention times and significantly higher selectivities atigh enantiomer concentrations. These are all aspects of utmostmportance considering the preparative separation of ketoprofennantiomers.
.4. Breakthrough experiments and simulation results
In order to test the equilibrium adsorption isotherms mea-ured, different saturation and regeneration curves were carriedut for the pure ethanol mobile phase and for the whole rangef feed concentrations (racemic mixtures of 10, 20, 30, and
ree mobile phases and different feed concentrations (CF1 + CF
2 ): (a) 0.1 (lineare are located at the upper-right corner; the separation regions for pure ethanol
d at the bottom-left corner.
3 urifica
4rdss
4
otbbbnc
SuwwseacmtavSermevd
Ffohe
bsitr
mrnolaCcrctcmdcd
Fparctf2ic
82 A.E. Ribeiro et al. / Separation and P
0 g/L). Fig. 7 shows the results obtained experimentally (byecovering and analyzing samples at different times) and pre-icted by the linear driving force model (Table 1). These figureshow a very reasonable agreement between experimental andimulation results in the whole concentration range.
.5. Simulated moving bed simulations
Chiral chromatographic separation processes are becomingf increasing importance in the development and the produc-ion of pharmaceutical drugs. Upon them, simulated movinged (SMB) technology is being applied by an increasing num-er of pharmaceutical companies. Its use at a production scale iseing considered as an alternative to the up to now leading tech-iques, such as enantioselective synthesis or diastereoisomericrystallization.
It is interesting to predict and compare the performance ofMB operation for the separation of ketoprofen enantiomerssing the three mobile phases in study. For this purpose, itas used the findings published by Morbidelli and co-workers,ho developed a complete design of the binary countercurrent
eparation processes by SMB chromatography in the frame ofquilibrium theory, assuming that mass transfer resistances andxial dispersion are negligible, and that the adsorption equilibriaan be described through a variable selectivity modified Lang-uir isotherm [9]. The SMB performance was evaluated for the
hree mobile phases by defining the complete separation regionsnd through the performance parameters of productivity and sol-ent consumption. A separation region is the area of possibleMB internal flow rates that allows 100% pure products (purextract, only containing the more retained enantiomer, and pureaffinate, only containing the less retained species). The perfor-
ance parameters of productivity and solvent consumption arevaluated at the vertex of each separation region. In fact, theertex is the point at the boundary of the separation region mostistant from the diagonal γ3 = γ2 (see Fig. 8) and represents the
ig. 9. Prediction of productivity of SMB operation under negligible mass trans-er resistances as a function of feed concentration: ratio between the productivitybtained with 100% ethanol and the one obtained with 20% ethanol/80% n-exane (solid line) and ratio between the productivity obtained with 100%thanol and the one obtained with 100% methanol (dashed line).
orrS
Fmtew
tion Technology 61 (2008) 375–383
est operating conditions in terms of system productivity andolvent consumption for a given feed concentration. For morenformation concerning SMB modeling and simulation, throughhe equilibrium theory and other more precise SMB models, seeeferences [9–14].
Fig. 8 shows the separation regions obtained for the threeobile phases at different feed concentrations. The separation
egions for 20% ethanol/80% n-hexane (at the upper-right cor-er) have operating conditions considerable different from thenes obtained for the pure alcohol mobile phases (at the bottom-eft corner), due to the higher retention times. This mobile phaselso leads to a stronger dependency on feed concentration.omparing the separation regions for the three mobile phases, itan be concluded that, for 20% ethanol/80% n-hexane, the sepa-ation region becomes quickly smaller with the increase of feedoncentration. This is a sign of stronger non-linear behavior ofhe adsorption process and a reason for lower productivities. Theomparison of the SMB performance for the two pure alcoholobile phases is also clear: both have similar operating con-
itions due to similar retention times, but pure ethanol presentsonsiderable better performances (bigger separation regions)ue to higher selectivity, as shown previously in Figs. 5 and 6.
Figs. 9 and 10 stress out the conclusions taken from Fig. 8.ig. 9 plots the ratios between the productivity obtained withure ethanol and the ones obtained with the other two solvents, asfunction of feed concentration. Fig. 10 plots the correspondent
atios of solvent consumptions. These simulation results alsolearly show that pure ethanol is the better choice for the separa-ion of ketoprofen enantiomers through SMB operation: at higheed concentrations (for example, 40 g/L of racemic mixture;0 g/L of each enantiomer) the productivity using pure ethanols three times the ones obtained with the other two solvents; theorrespondent solvent consumption is only 75% and 25% of the
ne needed with pure methanol and 20% ethanol/80% n-hexane,espectively. This last result also shows that, besides selectivity,etention times significantly influence solvent consumption inMB operation.ig. 10. Prediction of solvent consumption in SMB operation under negligibleass transfer resistances as a function of feed concentration: ratio between
he solvent consumption with 100% ethanol and the one obtained with 20%thanol/80% n-hexane (solid line) and ratio between the solvent consumptionith 100% ethanol and the one obtained with 100% methanol (dashed line).
urifica
5
scogut
fbuatTpp
A
(mP
Pm
R
[
[
[
[
A.E. Ribeiro et al. / Separation and P
. Conclusions
The optimization of preparative liquid chromatography andimulated moving bed for enantioseparation depends on thehoice of the proper mobile phase. In this choice, a high res-lution (or selectivity) of enantiomers should not be the onlyoal to be aimed. Other objectives, such as, to obtain high sol-bility of enantiomers and low retention times should also beaken into account.
This work shows that pure ethanol can be used for ketopro-en enantioseparation, instead of the mobile phase composedy 20% ethanol/80% n-hexane, which is the most commonlysed in analytical chiral HPLC. The pure ethanol mobile phasellows, simultaneously, high enantiomers solubilities, low reten-ion times and high selectivities at high feed concentrations.hese are all aspects of utmost importance at a preparative androduction scales, as it was shown for breakthrough adsorptionrocesses and SMB operation.
cknowledgements
Financial support by the Portuguese R&D foundation FCTFundacao para a Ciencia e a Tecnologia) and European Com-unity through FEDER (projects POCTI/EQU/38811/2001 andOCI/EQU/59738/2004), is gratefully acknowledged.
The authors wish to thank Simao P. Pinho (Bragancaolytechnic Institute) for the support on the solubilityeasurements.
eferences
[1] N.M. Maier, P. Franco, W. Lindner, Separation of enantiomers: needs,challenges, perspectives, J. Chromatogr. A 906 (2001) 3–33.
[
tion Technology 61 (2008) 375–383 383
[2] E. Yashima, Polysaccharide-based chiral stationary phases for high-performance liquid chromatographic enantioseparation, J. Chromatogr. A906 (2001) 105–125.
[3] K. Tachibana, A. Ohnishi, Reversed-phase liquid chromatography sepa-ration of enantiomers on polysaccharide type chiral stationary phases, J.Chromatogr. A 906 (2001) 127–154.
[4] E.R. Francotte, Enantioselective chromatography as a powerful alterna-tive for the preparation of drug enantiomers, J. Chromatogr. A 906 (2001)379–397.
[5] R.A. Granberg, A.C. Rasmunson, Solubility of paracetamol in pure sol-vents, J. Chem. Eng. Data 44 (1999) 1391–1395.
[6] L. Miller, C. Orihuela, R. Fronek, J. Murphy, Preparative chromatographicresolution of enantiomers using polar organic solvents with polysaccharidechiral stationary phases, J. Chromatogr. A 865 (1999) 211–226.
[7] S. Gracin, A.C. Rasmunson, Solubility of phenylacetic acid, p-hydroxyphenylacetic acid, p-aminophenylacetic acid, p-hydroxybenzoicacid, and ibuprofen in pure solvents, J. Chem. Eng. Data 47 (2002)1379–1383.
[8] M. Schulte, R. Ditz, R. Devant, J. Kinkle, F. Charton, Comparison of thespecific productivity of different chiral stationary phases used for simulatedmoving-bed chromatography, J. Chromatogr. A 769 (1997) 93–100.
[9] M. Mazzotti, G. Storti, M. Morbidelli, Optimal operation of simulated mov-ing bed units for nonlinear chromatographic separations, J. Chromatogr. A769 (1997) 3–24.
10] L.S. Pais, J.M. Loureiro, A.E. Rodrigues, Modeling, simulation and opera-tion of a simulated moving bed for continuous chromatographic separationof 1,1′-bi-2-naphthol enantiomers, J. Chromatogr. A 769 (1997) 25–35.
11] L.S. Pais, J.M. Loureiro, A.E. Rodrigues, Separation of enantiomers of achiral epoxide by simulated moving bed chromatography, J. Chromatogr.A 827 (1998) 215–233.
12] C. Migliorini, A. Gentilini, M. Mazzotti, M. Morbidelli, Design of simu-lated moving bed units under nonideal conditions, Ind. Eng. Chem. Res.38 (1999) 2400–2410.
13] G. Biressi, O. Ludemann-Hombourger, M. Mazzotti, R.M. Nicoud, M.
Morbidelli, Design and optimisation of a simulated moving bed unit: roleof deviations from equilibrium theory, J. Chromatogr. A 876 (2000) 3–15.14] L.S. Pais, V.G. Mata, A.E. Rodrigues, Simulated moving bed and relatedtechniques, in: G. Cox (Ed.), Preparative Enantioselective Chromatogra-phy, Blackwell Publishing, Oxford, UK, 2005, p. 176 (Chapter 7).